2 Sense monitors in one panel?

There is a lot of discussion about being able to have 2 sense monitor on one account but I’m curious about something else.
Has anyone put 2 monitors in the same panel?
I know you’d have to have 2 accounts but I think it would be interesting to take 2 and start them at the same time.
We all hear about how detections are different for reasons like homes being different and behavior.
I’d like to take both of those out of the equation and see what happens.

I would predict totally different results from the two monitors. Would it help the team to figure out some of the “why’s” of detections? I believe we are told “all house are different” but they really don’t know why.

1 Like

That might be interesting to see the variances of two different monitors, however I think we get why all houses are different. Just look at the monthly usage numbers we are posting from the reports we get emailed monthly. We have one member with a Bitcoin mining operation running MASSIVE always on numbers and other folks who’s usage just drips. I’ll bet the statistics on just the number of devices per house plugged in would be interesting to see.

1 Like

I would do a factory reset on my current monitor so the two started fresh.
It would probably turn into a side hobby trying to keep up with documenting and comparing the two.

The bitcoin miner didn’t even have to run his heat during winter because his machines kept them warm. I remember when that was posted, his use was shocking.

I too predict totally different activity, not sure what knowledge is to be had by this test though? I would assume Sense has tried this themselves, if they haven’t they should.

I’d be interested in the results none the less :slight_smile:

1 Like

My bet would be that two Senses similarly situated on the mains and the 220V breaker (same orientation on CTs and breaker) and with enough separation between the CTs to avoid mutual inductance effects between the two of them, would likely see similar detections over time. Ultimately, if both were well-calibrated, they would essentially see the “same” waveform with only minimal magnitude and phase differences. I say similar, because neural nets have non-linear activation functions throughout them, so occasionally small differences can blow up into big differences at the output.

2 Likes

@samheidie
I’m also unsure about what help it would be but I also don’t know everything about the detection process. I was thinking looking at it from their perspective they could compare why one did something different.
Even if it was not helpful to them, I would like to do it out of curiosity.

I thought about the CT’s and how much distance should be between them. My panel has the mains entering at the bottom and they run to the top, make a 180 and into the lugs.
The way I have the current monitor is one ct where the main enters and the other at the top at the lug. There is a foot and a half separation now. I would take a second monitor and do the same but on opposit legs. This would provide enough distance for all CT’s.

Your prediction is much different than what @samheidie and I are thinking.

I cheated a little though, I’ve had two monitors and they are quite different throughout the whole process.

  1. Different wattage tracking for each device (could be updated models)

  2. My first one loved lights, second has found one instead of a handful (could be newer firmware designed to disregard lights because of multiple lights being the same and causing user confusion)

  3. First monitor found almost all of the pumps and parts of my geothermal setup but took forever to find the main load (compressor itself), second monitor detected the compressor almost right away but hasn’t found any of the other pumps or components.

That’s a couple things off the top of my head. 1st and 2nd monitor are now installed for almost the same duration. To be clear, this is not a dual install, but a replacement unit.

1 Like

Assuming the two monitors were installed sequentially, one after the other, not concurrently ?

I would expect different results if they started at different point in time.

I would perforce factory reset on the current monitor and feed the 240 from the same breaker.
They would then have the exact start time

Yes Kevin, sequentially. Being there was no time lapse between installs I still find it hard to believe they’ve been so different.

I guess I’m more curious now than I was when I first read this thread :slight_smile:

1 Like

I’m still trying to parse when each Sense monitor was installed and when/whether the replaced monitor was deactivated…

From what I’m understanding he some problem with the first monitor. I’m assuming that his second monitor was installed and the same time the first was removed. That they did not ever run side by side.
I would have expected sense to not supply a second set of CT’s and Leads to prevent its use elsewhere or to disable the serial.

Unit 1 ran for 6 months by itself, deactivated, installed and activated new Sense box immediately (only replacing the monitor itself), 2nd replacement monitor running by itself for about 6 months now too.

That’s kind of what I thought. I would have expected different results / detections given different starting time points. The models would be different in each time period, due to regular training and model updates.

If it’s accurate, my monitor updates models every week or two (I have API exports to back this up when I remember.)

My thoughts would be that each monitor were on the same model for at least a while and the latter of the two monitors should have had better success if it was moving in the right direction.

Either way, this was never meant to be a scientific test, just sharing my experience in the department.

2 Likes

Wouldn’t be “scientific” when I do it either. Just really interesting.

@samheidie
If I were to have two monitors, each on a different model, couldn’t I run the one behind the other to catch up. Then reset both to factory and have the same starting point for both?

I guess if I were to do the test, I would have both units up and running and through the calibration stage.

From there I’d make sure both units are on the same exact firmware, if so, do an API call to each to check the model version, if same, factory reset both units at the same time.

The issue here is models are device specific and the new monitor wouldn’t have any model if it hasn’t detected anything. I believe if both units were at least on the EXACT firmware of each other and reset at the same time it would be a fair test.

2 Likes

Thank you for the response @samheidie
You explained what I was thinking much better than I did.
Checking some of the things you are talking about is honestly above my head.

1 Like